RESUMEN
According to the World Health Organization, birth-weight of less than 2,500 g is considered low birth-weight since below this value birth-specific infant mortality begins to rise rapidly. In Bangladesh, the prevalence of low birth-weight is unacceptably high. To screen low-birth-weight babies, simple anthropometric parameters can be used in rural areas where 80-90% of deliveries take place. A sample of 316 newborn singleton babies were studied in a government maternity hospital in Dhaka city to examine the birth-weight status of newborns and to identity the relationship between birth-weight and other anthropometric parameters of newborns. The mean birth-weight was 2,889 +/- 468 g, and 15.18% were low-birth-weight (<2,500 g) babies. All key anthropometric parameters of the newborns significantly correlated with one another (p<0.001). The best cut-offs for detecting low-birth-weight and normal-weight babies were <10 cm (odds ratio=17.4), <30.5 cm (odds ratio=25.0) and <33 cm (odds ratio=19.4), respectively, for mid-upper arm circumference (MUAC), chest circumference, and head circumference. The sensitivity and specificity were best for chest circumference (83.3% and 83.6% respectively). At lower cut-off points of <9 cm, <29.5 cm, and <32 cm, respectively, for MUAC, chest circumference, and head circumference, high-risk babies could be identified with a minimum number of false-positive cases. Chest circumference was the best detector of birth-weight with a correlation-coefficient of just above 0.84, followed by MUAC with a correlation-coefficient of just below 0.84. Based on the findings of the study, it is recommended to use <29.5 and 29.5 to <30.5 cm for chest circumference to identify 'high-risk' and 'at high-risk' newborns respectively.